Embodiment
As do not have special declaration, and in following present embodiment, inductance switch is meant master power switch pipe 44, the switch pin is meant SW end or SW pin (99 in the accompanying drawing just).The present invention is in order to explain and to compose a piece of writing conveniently, same technical characterictic has been adopted the different technologies term, but object its essence implication or representative or that point to is the same, is 39 in the accompanying drawing such as alleged primary inductance (being meant the inductance on elementary winding or the primary coil), elementary winding, primary coil; Also be the same such as secondary inductance (being meant the inductance on secondary winding or the secondary coil), secondary winding, secondary coil again, be in the accompanying drawing 40; It also is the same also having compole inductance (being meant the inductance on auxiliary winding, compole winding or the compole coil), auxiliary winding, compole winding, compole coil, is 41 in the accompanying drawing; Also having some conceptive different, but be last subordinate concept or equivalents or essence same concept, is exactly hereinafter inverse excitation type converter such as flyback converter, power supply changeover device; Modulator, adjuster, transformer all are meant 72 in the accompanying drawing.
Fig. 3 regulates the inverse excitation type converter 30 of inductive current with a comparison circuit and control loop for the present invention.Fig. 4 is the flow chart of inverse excitation type converter 30 method of works among Fig. 3, and it comprises step 31 to 35.The method is controlled the electric current of output by the peak current of regulating inverse excitation type converter 30 inductance.Inverse excitation type converter comprises 36, one external NPN triodes 37 of a transformer and a controller IC (being called for short " control IC ") 38.Transformer 36 comprises 39, one secondary winding 40 of an elementary winding (being called " primary inductance " again) and an auxiliary winding 41.Control IC 38 comprises the master power switch pipe 44 of 43, one inside of 42, one self-adaptive current limiters of an oscillator, 45 and gate driver circuits 46 of pulse-width modulation logic control circuit (being called for short " PWM logical circuit ").Self-adaptive current limiter 43 comprises 47, one control loops 48 of a comparison circuit and a pulse duration generator (being called for short " pulse width generator ") 49.
When 44 conductings of master power switch pipe, inductive current 50 begins to flow through elementary winding 39.Owing to rising on inductive current 50 slopes of flowing through elementary winding 39, produced an electromagnetic field, and when master power switch pipe 44 turn-offs, power transfer has been arrived secondary winding 40.The energy that is transferred to secondary winding 40 is just with output current (I
OUT) form output.In some applications, wish that inverse excitation type converter 30 can provide constant output current (I
OUT).Output current (I
OUT) relevant with three factors at least: (i) peak value of inductive current 50, the (ii) inductance value (L of primary inductance 39
PThereby) and (iii) 44 conductings of master power switch pipe flow through the frequency (f that the current ramp of primary inductance 39 rises
OSC).Inductance value (L with regard to primary inductance 39
P), because the variation of transformer 36 manufacture crafts causes its inductance value to depart from the rated value of its nominal, the therefore output current (I of other inverse excitation type converter
OUT) can change.For example, the diameter of inductance winding conducting wire is inequality, and perhaps the inconsistent actual inductance value of indivedual primary inductances that all can cause of the mode of Chan Raoing changes.In addition, thus the delay of signal transmission and use master power switch pipe 44 and control the peak current (I that ghost effect meeting that inductive current produces causes flowing through elementary winding 39
P) change.And transmission delay and technology, temperature is relevant with voltage.
Fig. 4 has described and has regulated the peak current (I that flows through elementary winding 39
P) a kind of method, although signal transmission delay and ghost effect and technology, the variation of temperature and voltage is relevant, it is constant that this way still can be kept the output current of inverse excitation type converter 30.In addition, can also be by regulating peak current (I
P) compensate the primary inductance 39 inductance value (L that cause owing to technique change
P) inconsistency.And, described by regulating frequency of oscillation (f
OSC) come compensating inductance value (L
P) a kind of method of inconsistency, master power switch pipe 44 is with this frequency of oscillation (f
OSC) conducting, make on inductive current 50 slopes of flowing through primary inductance 39 to rise.Therefore, by adjusting inductance peak current (I
P) and switching frequency (f
OSC) or only adjust the electric current (I that one can be kept output
OUT) be constant value.
The first step (step 31), self-adaptive current limiter 43 receiving feedback signals 51, the time that the electric current that primary inductance 39 is flow through in this signal reflection stops to increase.Comparison circuit 47 in the self-adaptive current limiter 43 and control loop 48 be the feedback signal 51 of reception oscillator 42 outputs all.The inductive current 50 that flows through primary inductance 39 stopped to rise on the slope in the time that stops to rise for the first time.Oscillator 42 uses auxiliary feedback signal 52 to produce feedback signal 51 and switching frequency signal 53.Auxiliary feedback signal 52 produces by auxiliary winding 41 terminal voltages.When the current ramp that flows through primary inductance 39 rose, magnetic field produced, and energy delivery is given auxiliary winding 41 and produced voltage at auxiliary winding 41 ends.
Second step (step 32), comparison circuit 47 receiving key signals 54, the speed that rises on inductive current 50 slopes of primary inductance 39 is flow through in this signal reflection.Switching signal 54 obtains from external NPN transistor emitter by the switch terminals (SW) of control IC 38.The inductive current 50 that rises on the slope in the primary inductance 39 flows through the switch terminals (SW) of NPN triode 37 and control IC 38.Though in Fig. 3, switching signal 54 is obtained by the NPN transistor emitter electric current that flows through master power switch pipe 44, but the implementation method that can use other produces switching signal 54, for example by connecting inductive reactance at the source of master power switch pipe 44 end or producing being connected resistance with the source end of the induction MOSFET of master power switch pipe 44 parallel connections.
In the 3rd step (step 33), comparison circuit 47 produces timing signal 55, and this signal reflection object time, this time is inductive current 50 reaches predefined restriction electric current with certain rate ramp rising time.
In the 4th step (step 34), control IC 38 produces inductor switch control signal 56, and this signal has certain pulsewidth.The grid of the master power switch pipe 44 that inductor switch control signal 56 control inductive currents 50 flow through.Gate driver circuit 46 uses " N-channel on " signal (" conducting of N raceway groove " signal is designated hereinafter simply as " Nchon signal ") 57 to produce inductor switch control signal 56.PWM logical circuit 45 uses the switching frequency signal 53 of oscillator 42 outputs and the pulse width signal (hereinafter to be referred as " pulse width signal ") 58 of pulse width generator 49 outputs to produce Nchon signal 57.Switching frequency signal 53 provides frequency for the pulse of inductor switch control signal 56 its generations, and pulse width signal 58 provides the pulse duration of inductor switch control signal 56.The time error signal 59 that pulse width generator 49 uses control loop 48 to produce produces pulse width signal 58.
The 5th step (step 35), the pulsewidth of self-adaptive current limiter 43 control inductor switch control signal 56 makes time of stopping to rise (flowing through the time that the inductive current 50 of primary inductance 39 stops to increase) and object time (inductive current 50 reaches the time of predefined restriction electric current) take place simultaneously for the first time.In the specific implementation circuit, the pulse duration of self-adaptive current limiter 43 control inductor switch control signal 56, but realize in the circuit pulsewidth or the Nchon signal 57 of self-adaptive current limiter 43 control pulse width signals 58 at another kind.By control pulse width signal 58, the pulse duration of any one signal among Nchon signal 57 and inductor switch control signal 56 threes, time that stops to rise for the first time and object time can be adjusted to simultaneously takes place.By adaptive control impuls width, the peak current (I of inductance
P) will be adjusted, thereby keep the output current (I of inverse excitation type converter 30
OUT) constant.
Fig. 5 is inverse excitation type converter 30 higher leveled block diagrams among Fig. 3.But inverse excitation type converter 30 is accurate power power-supply converter cheaply, and it is by elementary control, and output current is through adjustment.As shown in Figure 5, inverse excitation type converter 30 secondary control circuit and the optocoupler that do not comprise in the prior art to be comprised.Inverse excitation type converter 30 uses unique feedback from secondary to control output current and voltage, and this feedback is from the magnetic field coupling of auxiliary winding 41 and secondary winding 40.Except saving cost, owing to do not comprise secondary control circuit and optocoupler, peripheral parts number is reduced, thereby increases the stability of inverse excitation type converter 30.
There are two to influence inverse excitation type converter 30 output current factors of accuracy: (a) variation of the inductance value of transformer 36 its primary inductances 39 and (b) primary inductance 39 its peak current (I
P) inexactness that detects.Elementary winding 39 its actual inductance value (L
P) generally can change ± 20%.Peak current (the I of elementary winding 39
P) generally can not accurately detect, because the induction by current comparator in the control IC 38, PWM logical circuit and gate driver circuit have transmission signals to postpone, and the prime power switch turn-offs to have and postpones and as the ghost effect of the drain electrode of the MOSFET of prime power switch or as the ghost effect of the NPN transistor collector of prime power switch.In addition, its accuracy of detection of peak current is along with temperature, voltage, IC technology, the variation in PCB layout and the parasitic source relevant with the peripheral cell value and reducing.Inverse excitation type converter 30 is by making the operating frequency (f of master power switch pipe 44
OSC) variation and inductance value (L
P) be varied to the deviation that inverse ratio compensates the rated value of the actual value of primary inductance and its nominal.Inverse excitation type converter 30 uses self-adaptive current limiter 43 and control loop 48 to survey and the peak current of control primary inductance 39 comes compensating signal transmission delay and ghost effect, peak current is surveyed become simple.In addition, inverse excitation type converter 30 uses the switch structure of emitter electrode of elementary control for reducing cost.
Inverse excitation type converter 30 among Fig. 5 is respectively constant (peak value) current-mode and constant-voltage mode by two kinds of constant electric current and voltages of mode of operation output.Transformer 36 its elementary windings 39 have the Np number of turn, and secondary winding 40 has the Ns number of turn, and auxiliary winding 41 has the Na number of turn.Secondary resistance 60 in Fig. 5 is represented the resistive loss of transformer 36 its copper cash windings.Inverse excitation type converter 30 has secondary commutation device 61, output capacitance 62 and control IC 38.Control IC 38 is the peak-current mode PDM keyer.The primary power that control IC 38 starts is provided by resistance 63 and electric capacity 64.After inverse excitation type converter 30 was stablized, transformer 36 its auxiliary windings 41 provided energy by rectifier 65 for control IC 38.
Its routing pad 66 of the feedback end FB of the control IC 38 that transformer 36 is elementary receives reflection secondary winding 40 output voltage (V
OUT) signal.Terminal voltage (the V of auxiliary winding 41
AUX) 67 obtaining the auxiliary feedback signal 52 of its routing pad 66 ends of feedback end FB through resistance pressure-dividing networks, this resistance pressure-dividing network comprises first feedback resistance (R
FB1) 68 and second feedback resistance (R
FB2) 69.Auxiliary feedback signal 52 also is used to calculate the ON time and the actual slope rise time of primary inductance.
Comprise for example NPN triode 37 of external power control member in the specific implementation circuit of the inverse excitation type converter among Fig. 5, it is applied in the situation that needs higher power output or higher switching frequency.The base stage of NPN triode 37 is coupled to a diode 70 and a resistance 71.In low power applications, do not comprise external triode in the specific implementation circuit of inverse excitation type converter 30, MOSFET power switch pipe or current-sensing circuit, these all are integrated in the control IC 38.
Among Fig. 5, NPN triode 37 and control IC 38 collaborative works of adopting switch structure of emitter electrode.External NPN triode 37 is as the switch of elementary winding 39.In this structure, the emitter of the integrated external NPN triode 37 of drives in the control IC 38.In other realization circuit,, use external MOSFET to substitute NPN triode 37 as main switch in order further to increase power control ability and switching frequency.Usually, the frequency characteristic of triode is subjected to the restriction of base stage charge, and its high power characteristic is subjected to the restriction of base drive resistance.Therefore, NPN triode 37 is very suitable for the application that does not need very high power and switching frequency.
In the existing technology, adopt inductive reactance to come detection of primary inductance peak current, the method does not gear to actual circumstances, because flow through the electric current that the electric current of inductive reactance equals the NPN emitter junction, this electric current is made up of the inductive current of the reality that flows through collector electrode and the base current of NPN triode 37.Although use triode to increase complexity, and triode is because himself characteristic has current gain (Beta) and effect such as saturated, this effect can produce extra error term, but in application, still want to use the NPN triode to substitute MOSFET, this be because the cost of triode well below high-voltage MOSFET.Current gain and effect such as saturated are restive, and along with technology, temperature, the variation of voltage and peripheral cell value and sizable variation is arranged.
Fig. 6 is control IC 38 detail circuits figure.Control IC 38 comprises self-adaptive current limiter 43, and it is used for compensation and surveys primary inductance 39 peak current (I
P) time departure that produces.On the basis that does not influence performance, self-adaptive current limiter 43 is surveyed the error that has for the correction peak current low-cost solution is provided.
Although have the variation of each side in the system, self-adaptive current limiter 43 makes the peak current (I of primary inductance 39
P) constant.Control loop 48 is regulated the turn-off time of internal power MOSFET 44, makes the total slope rise time (T of primary inductance 39 its electric currents
RAMP) rise to predefined peak-limitation electric current (I corresponding to the primary inductance current ramp accurately
LIM) time.Total slope rise time (T
RAMP) comprising: (a) ON time of inner integrated master power switch pipe 44, (b) the base stage discharge time of NPN triode 37 and (c) 37 collector voltage rise time of NPN triode.Total slope rise time is adjusted to the twice that the current ramp that flows through elementary winding 39 rises to half required time of the peak current that is limited.In this example, use 2: 1 ratio, but in other realization circuit, can use other ratio.In the application of many reality, consider accuracy and concrete implementation method (for example coupling of device layout), its effect of 2: 1 ratio is fine.The ratio that other is suitable for example can during needing the occasion of special applications be used at 3: 1.Control loop 48 automatically impels 39 its actual current ramp rise time of primary inductance to equal the reference time.
Though the inconsistency of system can make the peak current difference of inductance, there are a lot of other application not need to keep very constant inductance peak current.In AC/DC power supply changeover device and the adapter, not needing to keep very a kind of application of constant inductance peak current is to protect it not enter the mistake state by restriction output current or power output.This application do not need as AC/DC off-line charger with output current adjust very accurate.
Inner integrated adjuster 72 provides supply voltage and reference voltage V for control IC 38
REFIn the specific implementation circuit, when circuit start, can produce 15 volts voltage V by resistance 63 and electric capacity 64
DD, auxiliary winding 41 and rectifier 65 can be kept voltage V behind the circuit start
DD, this voltage input adjuster 72, exporting 5 volts of supply voltages then is 43 power supplies of self-adaptive current limiter.Undervoltage lockout circuit 73 monitoring are the voltage V of control IC 38 power supplies
DD, as voltage V
DDWhen surpassing undervoltage lockout circuit 73 turn-on threshold voltage, control IC 38 operate as normal.In this example, undervoltage lockout circuit 73 turn-on threshold voltage are 19 volts, and it is 8 volts that undervoltage lockout circuit 73 turn-offs threshold voltage.If voltage V
DDBe reduced to undervoltage lockout circuit 73 and turn-off threshold voltage, control IC 38 will quit work.The reflected signal of the output voltage of transformer 36 its secondary winding 40 by auxiliary winding 41 and its routing pad 66 of feedback end FB, feeds back to control IC 38.The reference voltage V that auxiliary feedback signal 52 and adjuster 72 produce
REFCompare, output error signal, this error signal is amplified by preamplifier 74, through sampler 75 samplings, feeds back to PWM error amplifier 76, and this amplifier further amplifies error signal, and output is through twice amplified output signal 77.Resistance 78, electric capacity 79 and 80 is formed the internal compensation network of PWM error amplifier 76.The output signal 77 of PWM error amplifier 76 is input to error comparator 81, and this error comparator 81 is as the pulse-width modulation comparator of inverse excitation type converter 30 constant voltage modes.
Except preamplifier 74, auxiliary feedback signal 52 is input to oscillator 42 and frequency modulator (being called for short " FMOD ") 82 by its routing pad 66 of feedback end FB.The magnitude of voltage of the auxiliary feedback signal 52 of FMOD 82 induction, the bias current that output oscillator 42 is required.The bias current of FMOD 82 output changes with the variation of auxiliary feedback signal 52 magnitudes of voltage, and therefore when inverse excitation type converter 30 its output voltages variations, the frequency of oscillator can be along with adjustment, thereby it is constant to keep output current.Oscillator 42 comprises one and is used for detecting actual 39 current ramp rise time of elementary winding (T
RAMP) T
RAMPDetection circuit.T
RAMPDetection circuit is by the terminal voltage (V of auxiliary winding 41
AUX) 67 voltages that obtain through divider resistances 68 and 69 decide total slope rise time.The required frequency of oscillator 42 output pulse width modulation circuits is used to drive master power switch pipe 44.
The voltage of auxiliary feedback signal 52 is used for the reference voltage as oscillator 42 by the ratio decision of the inductance value of auxiliary winding 41 its inductance value and primary inductance 39 and secondary winding 40.Therefore, except peak current (I
P), frequency of oscillation (f
OSC) also compensate the variation of primary inductance 39 its inductance value.Except the physical circuit of Fig. 6, can also use the characteristic of other interchangeable structure optimization oscillator 42, come the variation of compensator transformer 36 its primary inductance values.
PWM logical circuit 45 utilizes two kinds of pulse width modulated waveform that mode producing is required: (a) when adjusting output voltage, adopt Converter Controlled by Current mode PWM and (b) when adjusting output current, adopt weekly phase self-adaptive current unrestricted model.Nchon signal 57 is input to gate driver circuit 46 by 45 outputs of PWM logical circuit.Gate driver circuit 46 is a MOSFET grid drive circuit relatively at a high speed.Gate driver circuit 46 outputting inductance switch controlling signals 56, this signal is input to the inside MOSFET 83 of master power switch pipe 44 and smaller scale.The inside MOSFET83 of smaller scale and resistance 84 are formed current-sensing circuit.Sensed current signal amplifies through induction by current amplifier 85, converts voltage signal then to.This voltage signal is compared with the output signal 77 of PWM error amplifier 76 by error comparator 81.Signal 86 is adjusted in error comparator 81 outputs, and it is used to set the ON time of master power switch pipe 44.Under the constant voltage mode of operation, when the output current of inverse excitation type converter 30 is lower than maximum export-restriction electric current, adjusts signal 86 and be used for constant voltage output and adjust.Under the constant current mode of operation, the adjustment function of output current is realized by self-adaptive current limiter 43, as output current (I
OUT) reach predefined peak-limitation electric current (I
LIM) time, the peak current (I of self-adaptive current limiter 43 restriction primary inductances 39
P).Self-adaptive current limiter 43 makes peak current value and temperature, input line voltage, and the variation of IC and peripheral cell value and the inconsistency of PCB layout are irrelevant.
The output signal 77 of PWM error amplifier 76 is input to line correction circuit 87, produces line corrected signal 88, and its value is proportional with output signal 77.Line corrected signal 88 is used to adjust the voltage of auxiliary feedback signal 52, compensates the loss of the output voltage that inverse excitation type converter 30 charger line series resistances cause.The online end of line resistance compensation technique provides rationally accurate fixed voltage, line terminal for inverse excitation type converter 30 be recharged and the equipment that is powered mobile phone or the portable media player end that links to each other for example.The loss of output voltage is that this voltage is reduced to the product of the output current of series resistance limited on the line and power supply because the voltage of load end has an IR voltage drop.Elementary control flyback power converter 30 is reflected to the feedback voltage of assisting winding 41 to obtain by the voltage by secondary winding 40 and adjusts output voltage (V
OUT), but this reflecting voltage does not comprise the IR voltage error that causes owing to limited line resistance.In the constant voltage mode of operation, the output current of the output of PWM error amplifier 76 and inverse excitation type converter 30 is proportional.Therefore output signal 77 can be used to produce line corrected signal 88, and the reference voltage input terminal that this line corrected signal 88 can be applied to feedback input end or preamplifier 74 comes compensating wire resistance.In the specific implementation circuit of Fig. 6, corrected signal is applied to the feedback input end of preamplifier 74, but in other interchangeable realization circuit, corrected signal can also simply be applied to reference voltage input terminal.
Oscillator 42 in Fig. 7 more detailed description control IC 38.Oscillator 42 comprises voltage comparator 89, delay element 90, T
RAMP91, three current sources 92,93 of detection circuit and 94 and oscillator timing electric capacity (C
OSC) 95.T
RAMPDetection circuit 91 is by auxiliary feedback signal total slope rise time of 52 decisions, and this feedback signal is auxiliary winding 41 its terminal voltage (V
AUX) 67 voltage signals that obtain through divider resistances 68 and 69.T
RAMPDetection circuit 91 output feedback signal (T
RAMP) 51.Feedback signal 51 is input to delay element 90, produces inhibit signal T
RAMPDInhibit signal T
RAMPDAfter feedback signal 51 produces, through T
D2Produce after time of delay.T
RAMPDetection circuit 91 comprises a current mirror 96 of being made up of P channel fet 97 and 98.When rising on inductive current 50 slopes of 44 conductings of master power switch pipe and primary inductance 39, oscillator 42 produces the VCO electric current I by current mirror 96
VCOThe VCO electric current I
VCOValue can be expressed as:
Wherein M is the gain of current mirror 96.In a kind of realization circuit, gain M is 1, I
VCOEqual the feedback current I that flows back to from its routing pad 66 of feedback end FB
FB
The charging current I that oscillator timing electric capacity 95 is produced by current source 92
OSCCharging.In this specific implementation circuit, oscillator timing electric capacity 95 is by current source 93 discharges, and the value of its discharging current is four times of charging current value.Because charging current source 92 is not closed when discharging current current source 93 is opened, so discharging current just becomes three times of charging current, as Fig. 9.When the master power switch pipe turn-offed, FMOD used with the proportional voltage signal of the voltage of assisting feedback signal 52 and produces bias current.This bias current bias current sources 92.Oscillator 42 is by 5 volts of power supply signal power supplies of adjuster 72 outputs.
Oscillator 42 is a RC oscillator that inside is integrated, output switching frequency signal 53, its frequency f
OSCCapacitance and oscillator resistor R by oscillator timing electric capacity 95
OSCResistance value determine.The resistance of oscillator can be expressed as R
OSC=V
FB/ I
OSC, V wherein
FB=V
OUTNa/Ns.The switching frequency signal 53 of oscillator 42 outputs is input to PWM logical circuit 45.PWM logical circuit 45 is by the switching frequency signal 53 of input and the pulse width signal 58 output Nchon signals 57 of pulse width generator 49 outputs.The frequency f of switching frequency signal 53
OSCThe frequency that 57 pulses of decision Nchon signal occur.
Fig. 8 is auxiliary winding 41 terminal voltage (V
AUX) 67, flow through the electric current (I of elementary winding 39
LP) and flow through the electric current (I of secondary winding
S) ideal waveform figure, wherein flow through the electric current (I of secondary winding
S) for flowing through the electric current of the secondary commutation device 61 that is operated in DCM.Master power switch pipe 44 turn-offs in the T2 time in T1 time conducting, in the conducting once more of T4 time.Therefore, the time between T1 and the T4 is a switch periods.Slope rise time (the T when time between T1 and the T2 is 44 conductings of master power switch pipe
RAMP).Time between T2 and the T4 is the time that master power switch pipe 44 turn-offs.Current waveform (I
S) shown in, in the T3 time, the electric current that flows through transformer 36 its secondary winding 40 is reduced to zero.
Feedback signal 51 (is voltage waveform T
RAMP) slope rise time of reflection primary inductance 39 reality, this time is by the terminal voltage (V of auxiliary winding 41
AUX) 67 detect by oscillator 42.The auxiliary feedback signal 52 of its routing pad 66 of feedback end FB provides the terminal voltage (V of auxiliary winding 41 for oscillator 42
AUX) 67 reflected signal.As shown in Figure 8, as terminal voltage (V
AUX) 67 become negative value and feedback signal 51 (T
RAMPWhen voltage) raising, primary inductance electric current (I
LP) begin to rise.As elementary inductive current (I
LP) arrive its peak value (I
P) time, oscillator 42 detects slope rise time T
RAMPFinish the voltage (V at auxiliary winding two ends
AUX) can raise rapidly.
The power output of inverse excitation type converter 30 only determines that by energy stored in the primary inductance 39 when the DCM shown in formula (2), this formula is ignored the loss that efficient is brought:
P
OUT=(V
OUT+V
D)·I
OUT=1/2·I
P 2·L
P·f
OSC (2)
V wherein
DBe the voltage drop at secondary commutation device 61 two ends, L
PBe the inductance value of elementary winding 39, I
PBe the peak current of elementary winding 39, f
OSCFrequency of oscillation for 42 settings of oscillator in the control IC 38.Ignore the loss of efficient, the output current of inverse excitation type converter 30 can be expressed as:
Work as I
OUTPeak current (the I that is limited less than elementary winding 39
P) time, the output voltage V of inverse excitation type converter 30
OUTBe the normal voltage of adjusting.The limit value of peak current preestablished before inverse excitation type converter 30 enters mode of operation.In the mode of operation of constant output electric current, when output current will surpass the constant output electric current of wanting, the output voltage V of inverse excitation type converter 30
OUTAdjustment voltage in the time of will be from operate as normal is reduced to zero.In order to keep I
OUTConstant, the switching frequency (f of oscillator 42
OSC) best and voltage (V
OUT+ V
D) proportional reducing, keep the peak current (I of elementary winding 39 simultaneously
P) constant.But because peak current (I
P) inconsistency, switching frequency (f
OSC) best and peak current (I
P) inversely proportional variation, thereby keep output current (I
OUT) constant.
Fig. 9 shows the inductance value (L of elementary winding 39
PAlthough) how by the feasible primary inductance (L of dynamic measurement
P) change switching frequency (f
OSCThereby) also keep output current (I along with changing
OUT) constant.Relevant described in Fig. 9 with following various formula.In addition, will introduce a kind of generation and elementary winding 39 inductance value (L below
P) be varied to the switching frequency (f that inverse ratio changes
OSC) method.Wherein, T
CHARGEExpression " charging interval ", T
DISCHExpression " discharge time ".
Produce switching frequency (f
OSC) final method by formula 11 statement.Portion waveshape figure among Fig. 9 is the desirable sequential chart of oscillator 42 in the control IC 38.Ramp voltage obtains the timing capacitor charge and discharge by current source.It in the oscillator 42 oscillator timing capacitor C
OSCThe electric current of charging is:
As Fig. 5, Na wherein is the number of turn of auxiliary winding 41, and Ns is the number of turn of secondary winding 40, R
FB1And R
FB2Be respectively the resistance value of feedback resistance 68 and 69, R
OSCBe resistance integrated in the oscillator 42, this resistance is used to produce bias current I
OSC, V
FBMagnitude of voltage for the auxiliary feedback signal 52 of its routing pad 66 of feedback end FB.Voltage V
FBObtain by two kinds of working conditions: (a) when master power switch pipe 44 turn-off and secondary winding 40 in electric current greater than zero the time, by terminal voltage (V
AUX) 67 obtain, its value equals (V
OUTNa/Ns) [R
FB2/ (R
FB1+ R
FB2)], (b) when 44 conductings of master power switch pipe, control IC 38 control voltage V
FB, make its value be approximately zero.Realize that at this as shown in Figure 9, selected oscillator timing capacitance discharges electric current is three times of charging current in circuit.In other realization circuit, can adopt other ratio.Attention discharging current source 93 in Fig. 7 is four times of charging current source 92, therefore obtains 3: 1 ratio.Oscillator frequency (f
OSC) by following formulae express:
V wherein
CO(V
COExpression timing capacitor C
VCOVoltage) be by another timing capacitor C
VCOWith charging current I
FBObtain.When 44 conductings of master power switch pipe, the voltage Be Controlled IC 38 of the auxiliary feedback signal 52 of its routing pad 66 of feedback end FB is pulled down to and is approximately zero.In addition, as shown in Figure 8, when 44 conductings of master power switch pipe, the voltage at auxiliary winding 41 two ends is negative value, and its value is:
Therefore,
With
Therefore, the output frequency of oscillator 42 can be by formula (4), and (5) and (8) draw, and its formula is as (9):
The weber of primary inductance can be expressed as:
V
IN·T
RAMP=L
P·I
P,(10)
The final expression formula that obtains the switching frequency of oscillator 42 generations is:
Or
Wherein K is the constant in the design.
Formula (12) illustrates the switching frequency (f that oscillator 42 produces
OSC) and voltage (V
OUT+ V
D) be directly proportional, with the inductance value (L of elementary winding 39
P) be inversely proportional to.Bringing formula (12) into formula (3) obtains:
I
OUT=1/2·K·I
P.(13)
Formula (13) illustrates the output current of inverse excitation type converter and the inductance value (L of elementary winding 39
P) irrelevant.Therefore, the adaptive control switching frequency f that is introduced
OSCMethod make f
OSCWith L
PBe inversely proportional to, effectively produce constant output current, its current value does not change with the variation of primary inductance value.
Formula (13) also illustrates inverse excitation type converter 30 accurate output current (I
OUT) can realize by the accurate peak current of setting primary inductance.Ordinary circumstance, the peak current (I of converter
P) not accurate the setting.For example, the peak current (I of transducer 25 of the prior art
P) use fixing reference voltage to set.Shown in Fig. 2 A (prior art), fixing reference voltage is obtained through external electric resistance partial pressure by bandgap voltage reference.Induction by current resistance (R
CS) 14 induced flows cross the electric current of primary inductance, are converted into voltage signal.When this voltage reaches reference voltage, trigger current is limited comparator, this comparator is the PWM logic reset, and closes main switch 12.The conventional method of this setting maximum primary inductive current itself has shortcoming.
Figure 10 is the work and the timing waveform of the control loop 48 of self-adaptive current limiter 43 in the control IC 38, wherein I
BASEExpression " base current of NPN triode 37 ", T
STARTExpression " time that inductive current begins to rise on the slope ", T
ILIM/2Expression " peak current that inductive current reaches setting is the time of ILIM/2 ".Nchon signal 57 is the conducting or the grid drive signal of turn-offing inner MOSFET, and this inner MOSFET is identical with master power switch pipe 44 functions.Switching signal 54 (waveform I
SW) hold routing pads 99 to arrive the electric current of inner integrated main switch mosfet 44 drain terminals for the emitter junction from external NPN triode 37 flows through control IC 38 its SW.Voltage waveform V
SWVoltage for SW end routing pad 99 these points.T
D1For Nchon signal 57 produces the back to the electric current I that flows through SW end routing pad 99
SWIn fact the time of delay between beginning to rise on the slope.This time of delay (T
D1) be the switching delay time of opening external NPN triode 37 (NPN 37 among Figure 10 just).Flow through the electric current (I of SW end routing pad 99
SW) form by two parts electric current: the actual current (I that (a) flows through primary inductance 39
LP), this electric current flows through the collector electrode of external NPN triode 37 and (b) base current of NPN triode 37.Base current is an offset current, and it makes the electric current (I of switching signal 54
SW) from the value of non-zero, as shown in figure 10, at (T time of delay
D1) when finishing, electric current (I
SW) rising that has rank to get over.Except the base current of NPN triode 37, also have other factor also can make the electric current (I that flows through primary inductance 39
LP) different with the electric current that flows through SW end routing pad 99, for example relevant ghost effect and transmission delay with main switch 44 drain terminals.
When oscillator 42 detects inductive current (I by auxiliary feedback signal 52
LP) when 50 beginnings rose on the slopes, oscillator 42 produced feedback signal (T
RAMP) 51.And flow through the inductive current (I of primary inductance 39
LP) 50 times that stop to rise, in Figure 10, use " FIRST TIME " note.When feedback signal 51 produces, the P channel fet is conducting, makes the current source (I that first is fixing
1) stored charge on first timing capacitor C 1.The slope accumulating rate of electric charge is dV on first timing capacitor C 1
C1/ dt=I/C1.Oscillator 42 is also exported T
RAMPDSignal, it is the inhibit signal of feedback signal 51.At first delay (T
D1) finish that the back produces second time of delay (T
D2) when finishing, oscillator 42 produces T
RAMPDSignal.Work as T
RAMPDWhen signal produced, second P channel fet be conducting, makes second fixing current source (I
2) stored charge on second timing capacitor C 2.In the specific implementation circuit of self-adaptive current limiter 43, the value of second timing capacitor C 2 is half of first timing capacitor C 1 in Fig. 6.In another alternative implementation method, the value of second timing capacitor C 2 is identical with the value of first timing capacitor C 1, second fixed current source (I
2) electric current that produces is first current source (I
1) twice of the electric current that produces.In these two kinds of realizing methods, the speed of the electric charge accumulation on second timing capacitor C 2 accurately be the twice of the speed that accumulates of the electric charge on first timing capacitor C 1.
As shown in figure 10, as inhibit signal T
RAMPDAfter the generation, electric charge begins accumulation on second timing capacitor C 2, through the ramp signal (I of base current compensation
SWCOMP) begin to rise, this ramp signal trace flow is crossed the switching signal 54 of SW end routing pad 99.The error part that base current in switching signal 54 electric currents causes is the ramp signal (I from compensating
SWCOMP) in by the place to go.Therefore, the ramp signal (I of compensation
SWCOMP) the reflection actual flow crosses the electric current (I of primary inductance 39 and NPN triode 37 collector electrodes
LP).
In the physical circuit of the self-adaptive current limiter 43 of Fig. 6, the ramp signal of compensation is eliminated the DC maladjustment part by the switching signal 54 of using coupling capacitance coupling SW end routing pad 99.At inhibit signal T
RAMPDBefore the generation, the electric charge on the coupling capacitance remains zero by a switch.Ramp signal (I when compensation
SWCOMP) reach a half of predefined elementary winding 39 its peak currents, the charging of second timing capacitor C 2 is suspended, the voltage of capacitor C 2 keeps.In concrete realization circuit, I
SWCOMPReach 1/2I
LIMTime be by the voltage (V of respective end relatively
SWCOMPAnd 1/2V
LIM) determine.The voltage that keeps on the capacitor C 2 is used for determining the ramp signal (I of compensation as reference voltage
SWCOMP) reach the precise time that flows through the peak current that elementary winding 39 limited.
The voltage that first timing capacitor C 1 is charged on first capacitor C 1 always reaches the reference voltage that is kept on second timing capacitor C 2.Timing signal 55 (also being called electric charge crossbar signal Tcx) is at the electric charge (V on first timing capacitor C 1
C1) reach second electric charge (V on the timing capacitor C 2
C2) time produces.When timing signal 55 produces, primary inductance electric current (I
LP) reach the peak-limitation electric current (I of restriction
LIM), this is because the charging rate of first timing capacitor C 1 is half of second timing capacitor C 2.Therefore, the time of generation timing signal 55 is exactly to reach the peak-limitation electric current (I that is limited
LIM) object time.
Then, the trailing edge of feedback signal 51 is compared with the rising edge of timing signal 55, and the time that the trailing edge of feedback signal 51 occurs is the time that the electric current of primary inductance 39 stops to rise, at this moment between primary inductance electric current (I
LP) arrive its peak value, and auxiliary winding both end voltage (V
AUX) raise rapidly.
As shown in Figure 6, PWM logical circuit 45 uses the pulse width signal 58 generation Nchon signals 57 that self-adaptive current limiter 43 produces.Therefore, the pulse duration of Nchon signal 57 is by 49 controls of the pulse width generator in the self-adaptive current limiter 43.Pulse width signal 58 is compared the trailing edge of feedback signal 51 by the control loop 48 that uses the delay lock ring structure with the rising edge of timing signal 55.The control loop 48 of DLL type comprises a phase discriminator, and its trailing edge in feedback signal 51 shifts to an earlier date when timing signal 55 desired rising edges arrive, and produces the down pulse, expansion feedback signal trailing edge, thereby the duty ratio of increase signal Nchon signal 57.Increase its duty ratio by the trailing edge that postpones Nchon signal 57, thereby flow through the peak current (I of primary inductance 39 in next switch periods increase
P).In like manner, when the trailing edge of feedback signal 51 lagged behind timing signal 55 desired rising edges and arrives, the phase discriminator output up pulse in the control loop 48, up pulse be by the trailing edge of feedback signal in advance, thereby reduce the duty ratio of signal Nchon signal 57.Reduce duty ratio by the trailing edge that shifts to an earlier date signal Nchon signal 57, thereby reduce to flow through the peak current (I of primary inductance 39 in next switch periods
P).Therefore control loop 48 is kept the peak current (I of primary inductance 39
P) be predefined peak-limitation electric current I
LIM
As shown in figure 10, second time of delay (T
D2) be feedback signal 51 and inhibit signal T
RAMPDBetween delay, as long as second time of delay is less than the ramp signal (I of compensation
SWCOMP) reach 1/2I
LIMThe needed time, its length does not influence the ramp signal (I of compensation
SWCOMP) reach the peak-limitation electric current (I of predefined elementary winding 39
LIM) half time.This is true, because the voltage (V on second timing capacitor C 2
C2) determine the electric charge on first timing capacitor C 1 to reach second reference voltage (V on the timing capacitor C 2
C2) time, the ramp signal of this time and compensation arrives 1/2I
LIMPrecise time corresponding.
Control loop 48 in the self-adaptive current limiter 43 is adjusted timing signals 55, and when making the trailing edge of the rising edge of timing signal 55 and feedback signal 51 arrive simultaneously, the peak current of primary inductance 39 equals predefined peak-limitation electric current.Control loop 48 makes peak current (I
P) consistent with predefined peak-limitation electric current, and be not subjected to the variation of input line voltage, temperature, technology to a great extent, the inconsistent influence of the tolerance variations of element and PCB fabric swatch.
Available another kind of method is further set forth, and inner integrated main switch mosfet 44 is in T1 conducting in the time, and its time is the ramp signal (I of compensation
SWCOMP) arrival 1/2I
LIMTime add the time (T of the preceding paragraph variable-widthization
WIDTH).The time of this variable-widthization is the time of the pulse width variations of Nchon signal 57.Frequency of oscillation (the f that master power switch pipe 44 produces at oscillator 42
OSC) each conducting when arriving in cycle, at (T1+T
WIDTH) shutoff, wherein T when finishing
WIDTHAdjust by control loop 48, the slope rise time that the result equals to wish total slope rise time, thus it is constant to keep output current.
Figure 11 is the self-adaptive current limiter 43 more detailed description figure in the control IC 38.Self-adaptive current limiter 43 comprises comparison circuit 47, control loop 48 and pulse width generator 49.Pulse width generator 49 comprises ono shot pulse generator 100, and it produces a pulse when Nchon signal 57 produces suitable pulse duration.Control loop 48 comprises 101, one charge pumps 102 of a phase discriminator and loop filter 103.Control loop 48 is similar to delay phase-locked loop (DLL), and makes feedback signal 51 and timing signal 55 synchronous.Phase discriminator 101 comprises two D-flip-flops (d type flip flop) 104 and 105 and NAND (NAND gate) 106.Charge pump 102 comprises two switches 107,108, and two current sources 109,110.Loop filter 103 comprises a resistance 111 and an electric capacity 112, and time error signal 59 produces voltage signal V through its filtering
FILTEROno shot pulse generator 100 in the pulse width generator 49 is at the ramp signal (I when compensation
SWCOMP) reach reference current 1/2I
LIMIn time, reset, and when single triggering timing device 113 timing finish, is cleared.Ono shot pulse finishes the back a time period and produces the voltage signal V that this time period and time error signal 59 produce through filtering
FILTERBe inversely proportional to, be directly proportional to the time difference between timing signal 55 rising edges with feedback signal 51 trailing edges.
Self-adaptive current limiter 43 also comprises first timing electric capacity (C1) 114, second timing electric capacity (C2) 115, three timing bias current sources 116-118, first comparator 119, second comparator 120, two P channel fets (field-effect transistor) 121-122,123, one electric capacity 124 of a N channel fet and an inductive reactance (R
SENSE) 125.The value of first timing electric capacity (C1) 114 is the twice of second timing electric capacity (C2), 115 values.
Electric current (I when elementary winding 39
LP) begin to rise and feedback signal 51 when having produced, P channel fet 121 turn-offs, and timing bias current sources 117 begins first timing electric capacity (C1) 114 chargings.Therefore, as Figure 10, the electric charge (V on first timing capacitor C 1
C1) begin to rise on the slope.Second time of delay (T
D2) after the end, inhibit signal T
RAMPDProduce, P channel fet 122 turn-offs, thereby timing bias current sources 118 begins 115 chargings of second timing electric capacity.The rate of rise of voltage is the twice of first timing electric capacity 114 on second timing electric capacity 115, because the capacitance of second timing electric capacity 115 is first timing electric capacity 114 half.
As inhibit signal T
RAMPDDuring generation, N channel fet 123 turn-offs, the ramp signal (I of base current compensation
SWCOMP) being input to the positive input of first comparator 119, the ramp signal of this compensation is removed switching signals 54 electric current (I by electric capacity 124
SW) in the DC maladjustment electric current that causes by external NPN triode 37 base currents and produce.Then, first comparator 119 is with voltage signal (V
SWCOMP) and 1/2V
LIMSignal compares, wherein V
SWCOMPRamp signal (the I of corresponding compensation
SWCOMP), 1/2V
LIMSignal is produced by timing bias current sources 116 and resistance 126, its value and reference current 1/2I
LIMCorresponding.In other realization circuit, adopt have can faradic FET current comparator replace first voltage comparator 119, can be directly with the ramp signal (I that compensates
SWCOMP) and reference current 1/2I
LIMCompare.Ramp signal (I when compensation
SWCOMP) reach reference current 1/2I
LIM, first comparator 119 produces energizing signal, and this signal turn-offs P channel fet 127, therefore turn-offs timing bias current sources 118.When timing bias current sources 118 is turn-offed, the electric charge (V on second timing electric capacity 115
C2) remain unchanged.And the electric charge (V on first timing electric capacity 114
C1) rise with half charge rate of second timing electric capacity 115.Second comparator 120 is with the electric charge (V that keeps on the electric charge that continues to increase on first timing electric capacity 114 and second the timing electric capacity 115
C2) compare.As the electric charge (V that continues to rise
C1) reach on second the timing electric capacity 115 electric charge (the V that keeps
C2), the object time reaches, and second comparator 120 produces timing signal 55.Phase discriminator 101 is with the rising edge of the timing signal 55 electric current (I as primary inductance 39
LP) equal the predefined time of flowing through the peak-limitation electric current of primary inductance 39.
In the specific implementation circuit of Figure 11, the relative value of first and second timing electric capacity 114-115 is used to produce timing signal 55 its correct sequential.Thereby other circuit structure also can use and obtain correct sequential.For example, can use onesize electric capacity, simultaneously the electric current that produces of first timing bias current sources 117 is half of the electric current that produces of second timing bias current sources 118.Perhaps when timing electric capacity and current source were all identical, second comparator 120 was at the voltage (V that continues to rise
C1) be the voltage (V that keeps
C2) twice the time, just produce timing signal 55.
In the specific implementation circuit of Figure 11, by the voltage signal V through filtering of control loop 48 generations
FILTER,, be used for reflecting the time difference between the rising edge of the trailing edge of feedback signal 51 and timing signal 55 as time error signal 59.As the ramp signal I that is compensated
SWCOMPRise to predefined fixing reference current value 1/2I
LIMThe time, circuit produces the rising edge of timing signal 55.In another kind of specific implementation circuit, through filtered voltage signal V
FILTERBe used to adjust the reference current 1/2I that produces by timing bias current sources 116 and resistance 126
LIMLike this, the voltage on second timing electric capacity 115 will reach reference voltage simultaneously with the voltage on first timing electric capacity 114.In a kind of like this specific implementation circuit, when the trailing edge of feedback signal 51 in advance when the rising edge of timing signal 55 arrives, reference current value 1/2I
LIMWill increase, thereby expression needs to increase the peak current (I of primary inductance
P); Correspondingly, when the trailing edge of feedback signal 51 lagged behind the rising edge arrival of timing signal 55, reference current value will reduce, thereby expression need reduce the peak current (I of primary inductance
P).
In another kind of specific implementation circuit, the switching frequency (f of oscillator 42
OSC) will adjust according to time error signal 59, thereby make inverse excitation type converter 30 produce constant output current I
OUTAs shown in Equation (5), for given oscillator timing capacitor C
OSC, can be by adjusting the charging current I of oscillator
OscAdjust its switching frequency.And I
OscValue can adjust by the resistance R osc that changes in the chip internal oscillator.Above-mentioned formula (3) shows: I
OUTSwitching frequency (f with oscillator 42
OSC) proportional.Therefore, adjust switching frequency (f according to time error signal 59
OSC), thereby at the peak current (I of primary inductance
P) under the situation about changing, keep output current I
OUTConstant, wherein time error signal 59 is to produce according to the delay between the trailing edge of object time and feedback signal 51.In formula (3), it may be noted that output current I
OUTPeak current (I with primary inductance 39
P) square proportional, therefore, in order to keep output current (I
OUT) constant, switching frequency (f
OSC) must with square the adjusting inversely of peak current (Ip).
Realizing in the circuit more specifically, in order to keep output current (I
OUT) constant, PWM error amplifier 76 is adaptively adjusted its output area according to time error signal 59, and time error signal 59 is to produce according to the delay between object time and feedback signal 51 trailing edges.When inverse excitation type converter 30 works in normal constant voltage mode, the magnitude of voltage of the output signal 77 of PWM error amplifier 76 and output current (I
OUT) proportional.In addition, under constant voltage mode, the ON time of master power switch pipe 44, it is by the time signal T among Fig. 8
RAMPExpression is to be controlled jointly by the output voltage of induction by current amplifier 85 and the output signal 77 of PWM error amplifier 76.When output current increases, the also corresponding increase of the magnitude of voltage of the output signal 77 of PWM error amplifier 76, thus keep the constant of output voltage.
Usually, conducting when master power switch pipe 44 arrived in each clock cycle, and the output voltage signal of induction by current amplifier 85 will be along with primary inductance electric current (I
LP) rise on proportional slope, and primary inductance electric current (I
LP) ramp-up rate be dl/dt=V
P/ L
P, wherein Vp is the voltage at primary inductance two ends.When the output voltage signal of induction by current amplifier 85 reached the output signal 77 of PWM error amplifier 76, master power switch pipe 44 turn-offed.Therefore the adjustment signal 86 by 81 outputs of clamper error comparator, thereby with the peak current (I of primary inductance
P) be limited in certain maximum; By the clamp voltage of adjustment signal 86, thus the peak current (I to being limited
P) control.To adjust clamp voltage adaptively by the time error signal 59 that control loop 48 produces, thereby keep output current (I
OUT) constant.In this realization circuit, no matter inverse excitation type converter 30 adjusting outputs obtain constant voltage still is constant electric current, the time that the master power switch pipe turn-offs all is to decide the time that voltage signal by 85 outputs of induction by current amplifier reaches the output signal 77 of PWM error amplifier 76 all the time.When inverse excitation type converter 30 works in following time of limit of constant voltage mode, the voltage of output signal 77 will be in being lower than the normal range (NR) of clamp voltage, and under constant current mode, the voltage of output signal 77 is clamped at maximum and comes peak limiting electric current (I
P). under constant current mode, thereby control loop 48 will be adjusted clamp voltage value control T adaptively
RAMPTime, keep output current (I
OUT) constant.
For better elaboration, the present invention uses some concrete realization circuit to be described, but the present invention is not limited in the implementation method of being mentioned.For example, other specific implementation circuit can use adaptive primary inductance compensation, rather than adopts self adaptation primary inductance peak current limit.And, adopting the structure of penetrating grade switch for high pressure NPN triode external among Fig. 5 37, other specific implementation circuit can directly drive elementary winding 39 by the inner integrated high-voltage circuit breaker of control IC.In addition, in order further to improve the switching frequency of power handling capability and inverse excitation type converter 30, can use a MOSFET to replace triode as external switch.
Figure 12 is another alternative realization circuit of PWM controller IC 128.This controller IC 128 do not comprise inner main switch mosfet, be used for induction by current small scale MOSFET and be coupled to induction by current resistance on the induction by current MOSFET, just do not comprise master power switch pipe 44, inner MOSFET 83 and resistance 84 among Fig. 6.In this realization circuit, the current driving ability of its gate driver circuit 46 can better be controlled for large-sized MOSFET.
Figure 13 is another alternative realization circuit of the inverse excitation type converter 30 of the controller IC 128 among use Figure 12.The realization circuit of this inverse excitation type converter 30 comprises an outside MOSFET 129 and an induction by current resistance 130.
Figure 14 is for adopting the inverse excitation type converter 30 of controller IC 38 its integrated circuit encapsulation 131.This controller IC 38 only uses auxiliary feedback signal 52 to control the output current and the output voltage of inverse excitation type converter as feedback signal, so its integrated circuit encapsulation only has four ports.The increase of the port number of encapsulation can make the corresponding raising of chip cost.Therefore, compare more than four controller IC, adopt with integrated circuit encapsulation 131 controller ICs that encapsulate 38 costs lower with encapsulation port.Integrated circuit encapsulation 131 has only four ports: 133, one power ports 134 of 132, one feedback ports of a switch port and a ground port one 35.In the specific implementation circuit of Figure 14, switch port 132 links to each other with SW end routing pad 99 by joint line 136.Switching signal 54 is received by switch port 132 and is transferred to SW end routing pad 99 by joint line 136.When the form of encapsulation not simultaneously, the form of switch port 132 is also different.For quad flat package, switch port 132 is pins; For Organic Land Grid Array (LGA), switch port 132 is terminal pads; For pin type grid array (GPA), switch port 132 is pin type pins; For dual in-line package (DIP) or single in-line packages, switch port 132 is pins.When the encapsulation 131 of integrated circuit adopted baii grid array packaged type and control IC 38 to adopt the mode of flip chip bonding to encapsulate, switch port 132 was not to be connected on the SW end routing pad 99 by joint line.When adopting baii grid array as encapsulation scheme, there is individual cushion pad at 99 places at SW end routing pad, and switch port 132 is connected to cushion pad by a wire ball.In different encapsulation implementations, feedback port 133, power port 134 and ground port one 35 can be that of baii grid array is connected ball equally, a pin of quad flat package, an or terminal pad of Organic Land Grid Array (LGA), an or pin type pin of pin type grid array (GPA), an or pin of dual in-line package (DIP) or single in-line packages.Be connected to by joint line 137 in the embodiment of feedback port 133 at its routing pad 66 of feedback end FB, control IC 38 receives one by feedback port 133 can reflect secondary winding 40 output voltage (V
OUT) signal.Auxiliary feedback signal 52 is received by feedback port 133 and is transferred to pad FB by joint line 137.
Though above-mentioned PWM logical circuit 45 adopts pulse width modulation to produce Nchon signal 57 and inductor switch control signal 56, can use the variable-frequency pulse modulation to replace the PWM of constant frequency.In this realization circuit, the mode of the pulse modulation (PFM) that the chip frequency of utilization changes produces Nchon signal 57 and inductor switch control signal 56.
Thereby various improvement are adjusted and its characteristic synthesis of above-mentioned implementation method all belongs within this scope of invention.